0:00:00.000,0:00:13.149 33c3 pre-roll music 0:00:13.159,0:00:15.269 Herald: Err ... 0:00:15.289,0:00:17.820 H: ... a talk would be good, right? 0:00:18.290,0:00:24.529 applause 0:00:26.169,0:00:27.330 Do you want to give a talk? 0:00:27.340,0:00:31.390 Toni: Aah, it’s a little early[br]but I’ll try. 0:00:31.390,0:00:36.360 Herald: Okay, guys, well, I found someone[br]who’s willing to give a talk! 0:00:36.430,0:00:41.820 laughter and applause 0:00:42.430,0:00:47.010 That is most excellent.[br]So, if you ever asked yourself, 0:00:47.770,0:00:53.120 I’ve got this big regime and[br]I’m rolling out internet censorship, 0:00:53.120,0:00:56.449 what does my economy do? 0:00:56.449,0:00:59.449 There are people in here[br]asking that question, right? 0:00:59.449,0:01:02.750 There’s always someone at Congress[br]who’s asking some question. 0:01:02.760,0:01:09.310 Well, you came to the right place,[br]and as part of her PhD thesis work 0:01:09.320,0:01:15.030 Toni is going answer that question,[br]hopefully, to a satisfactory point. 0:01:15.030,0:01:17.570 Please give a warm round of applause![br]applause 0:01:17.580,0:01:24.180 Toni![br]ongoing applause 0:01:24.180,0:01:27.520 Toni: Okay, thanks everyone for being[br]here, I hope you can all hear me 0:01:27.540,0:01:32.590 correctly. And I’m glad to be here[br]and to be presenting 0:01:32.590,0:01:36.109 some part of my thesis to day.[br]Now, this is ongoing work 0:01:36.109,0:01:39.520 so I’m really grateful for any kind of feedback[br]that you guys would have 0:01:39.520,0:01:43.300 and I’m really only presenting this[br]as kind of a first try, 0:01:43.300,0:01:46.840 because when I looked at the topic[br]of internet censorship 0:01:46.840,0:01:51.549 and what that could mean for an economy,[br]I really didn’t find anything academic 0:01:51.549,0:01:56.280 and I was quite surprised: it seemed[br]like a very obvious question to me, 0:01:56.280,0:02:00.979 because I was looking mostly[br]at China at the beginning. 0:02:00.979,0:02:04.740 And I read a lot of newspaper articles[br]and I talked to a lot of businessmen 0:02:04.740,0:02:08.060 who told me: “Well, doing business[br]in China is very difficult” 0:02:08.060,0:02:11.020 and I think China is really[br]holding itself back by having 0:02:11.020,0:02:15.400 this big censorship thing going. 0:02:15.400,0:02:20.670 But no one really looked into[br]how it is holding itself back 0:02:20.670,0:02:23.860 or if it is even holding itself back. 0:02:23.860,0:02:26.940 So there is really[br]very, very little research. 0:02:26.940,0:02:32.050 And we don’t even have an agreement among[br]economists or business studies people 0:02:32.050,0:02:36.420 about what impact the internet has[br]on the economy. So if you want to ask: 0:02:36.420,0:02:39.890 “So what does internet censorship do[br]to an economy?” it seems pretty obvious 0:02:39.890,0:02:44.830 to first ask: “What does the internet do to[br]an economy?” and we don’t even know that. 0:02:44.830,0:02:47.990 That was quite surprising to me and I’m[br]going to be talking about the reasons 0:02:47.990,0:02:53.310 for that a little bit later on. But in[br]general, I was thinking of a research 0:02:53.310,0:02:58.460 question to ask which for me is: “Does[br]internet censorship reduce economic welfare?” 0:02:58.460,0:03:03.030 Now, not all of you are economists,[br]so some of you might think of welfare 0:03:03.030,0:03:08.200 more as the transfer payments[br]that a state gives to its poorer people. 0:03:08.200,0:03:12.800 But for economists, economic welfare[br]is defined as the consumer 0:03:12.800,0:03:18.130 and producer surplus. So basically, the[br]difference between what something costs 0:03:18.130,0:03:21.760 and what you can sell it for[br]is the producer surplus. 0:03:21.760,0:03:25.160 The difference between[br]what you would be willing to pay 0:03:25.160,0:03:28.250 and what you’re actually paying[br]is your consumer surplus. 0:03:28.250,0:03:32.360 Now let’s assume I have a laptop[br]and I bought this. 0:03:32.360,0:03:35.910 And I would have been willing to pay[br]€ 1500 for this laptop because 0:03:35.910,0:03:39.860 I think it’s a very good product,[br]it’s by Lenovo that makes good laptops. 0:03:39.860,0:03:44.260 But actually I got it for like €800[br]or €900. That would mean 0:03:44.260,0:03:49.410 my personal consumer surplus[br]is something like €600 or €700. 0:03:49.410,0:03:52.910 And if we add up everyone’s[br]individual consumer surplus 0:03:52.910,0:03:58.840 we get the economic welfare surplus. 0:03:58.840,0:04:02.660 So first, I was trying to figure out[br]what does the internet mean 0:04:02.660,0:04:07.630 for the economy. And I’ve said that there[br]is really no good agreement on that. 0:04:07.630,0:04:12.330 Now, a very crude measure that I found is[br]how much does "the Internet economy" 0:04:12.330,0:04:17.780 contribute to GDP?[br]Now, what is "the internet economy"? 0:04:17.780,0:04:22.280 It wasn’t very clear in the research[br]that I’ve read. It seems to be sort of 0:04:22.280,0:04:27.620 online retail, and possibly some other[br]internet-enabled services? 0:04:27.620,0:04:31.130 Possibly but not necessarily[br]internet advertisement revenue 0:04:31.130,0:04:36.140 is reflected in this. But because it was[br]BCG, which is a big consulting agency 0:04:36.140,0:04:40.870 that basically published this research[br]they weren’t very diligent about 0:04:40.870,0:04:45.670 their methods, basically.[br]So we can see, well it seems that the UK 0:04:45.670,0:04:49.720 has a pretty big part of internet economy[br]as part of GDP. 0:04:49.720,0:04:53.760 That’s probably mostly because of[br]online retail which is bigger in the UK 0:04:53.760,0:04:57.310 than in most other countries we look at.[br]And we see that there is 0:04:57.310,0:05:01.680 a small difference between[br]developed and developing market averages 0:05:01.680,0:05:06.980 when looking only at the G20 countries.[br]But this seems like a very 0:05:06.980,0:05:10.330 dissatisfactory answer because first[br]of all, I don’t know the methods, 0:05:10.330,0:05:12.870 so I can’t really say[br]whether this is actually good. 0:05:12.870,0:05:16.350 And secondly, GDP is actually[br]not a good measure 0:05:16.350,0:05:20.490 for what we are trying to measure because[br]a lot of the stuff that the internet creates, 0:05:20.490,0:05:25.830 a lot of the value the internet creates[br]isn’t captured by GDP at all. 0:05:25.830,0:05:30.310 One example is free online courses.[br]Most of the online courses you can take 0:05:30.310,0:05:34.290 on the web are actually free.[br]And most of them are not ad-enabled. 0:05:34.290,0:05:40.690 So most of them don’t really have[br]advertisements in the general sense. 0:05:40.690,0:05:46.380 So classical economics basically says:[br]“Well, they don’t really create any value.” 0:05:46.380,0:05:48.650 But if you’ve ever taken[br]one of these online courses, 0:05:48.650,0:05:51.380 and maybe you’ve been lucky[br]and took a good one 0:05:51.380,0:05:54.100 you would actually… I would say that[br]some of the courses I took, 0:05:54.100,0:05:57.780 they created some value for me.[br]So one of the ways to look at this 0:05:57.780,0:06:02.660 is actually to think about time as[br]something that has opportunity cost. 0:06:02.660,0:06:06.180 So if I’m spending my time doing this[br]online course I’m not spending it 0:06:06.180,0:06:11.030 e.g. earning money. I’m also not[br]spending it doing something leisurely 0:06:11.030,0:06:17.749 that is fun for me.[br]And these guys, Brynjolfsson 0:06:17.749,0:06:21.050 – I’m sorry I don’t know[br]how to pronounce it exactly, 0:06:21.050,0:06:26.110 he sounds Swedish, possibly –[br]and ohh, in 2012 0:06:26.110,0:06:33.020 they tried to get an idea of[br]how much consumer surplus 0:06:33.020,0:06:38.950 these online courses actually create.[br]Which isn’t at all 0:06:38.950,0:06:44.360 reflected in the GDP.[br]And you see that in some models 0:06:44.360,0:06:49.550 it would be 5% of GDP[br]for these online courses alone. 0:06:49.550,0:06:56.750 Even if we take their more... most conservative[br]model which is $4.18 billion 0:06:56.750,0:06:59.729 on average for the years 2008-2011, 0:06:59.729,0:07:03.890 that’s still a pretty significant chunk[br]of economic welfare 0:07:03.890,0:07:07.780 that’s somehow being created[br]that is not reflected in GDP 0:07:07.780,0:07:11.770 because GDP is only stuff[br]that you actually pay money for. 0:07:11.770,0:07:14.919 Another example that we[br]might think of is Wikipedia. 0:07:14.919,0:07:19.160 Now Wikipedia has a certain cost of[br]operating: obviously the servers and stuff. 0:07:19.160,0:07:22.990 But because most people contributing[br]to Wikipedia are actually volunteers 0:07:22.990,0:07:25.910 the cost of operating[br]does not really reflect 0:07:25.910,0:07:30.250 the true value Wikipedia creates.[br]And one of the… 0:07:30.250,0:07:32.860 even if you don’t want to say…[br]even if you don’t agree 0:07:32.860,0:07:37.401 that time has opportunity cost, what[br]about the money that you don’t spend 0:07:37.401,0:07:43.800 on encyclopedias? How many of you guys[br]have encyclopedias at home? 0:07:43.800,0:07:46.210 OK, that’s more than I expected! 0:07:46.210,0:07:49.260 How many of you guys have[br]recent encyclopedias at home? 0:07:49.260,0:07:53.580 That’s a little less, this is kind of more[br]what I was expecting. 0:07:53.580,0:07:57.880 And now, my family also… we also have[br]an encyclopedia at home. 0:07:57.880,0:08:02.720 I think it’s from 1985 or something.[br]And before this encyclopedia 0:08:02.720,0:08:06.210 we would regularly update an encyclopedia,[br]we would regularly go out and buy 0:08:06.210,0:08:09.840 a new encyclopedia because[br]knowledge changed, obviously. 0:08:09.840,0:08:14.410 But ever since probably 1990,[br]we just didn’t bother. 0:08:14.410,0:08:20.630 So, assuming an encyclopedia might,[br]like a physical book, might cost €100. 0:08:20.630,0:08:24.400 And assuming sort of 2/3[br]of all households in Germany 0:08:24.400,0:08:28.100 have had an encyclopedia at one point. 0:08:28.100,0:08:31.710 We’re looking at 13 million households[br]at this point. 0:08:31.710,0:08:35.630 Now you don’t buy an encyclopedia[br]every year but you might buy it 0:08:35.630,0:08:40.679 every ten years. So in order to simplify[br]this we can say, every year 0:08:40.679,0:08:46.010 1.3 million households buy[br]an encyclopedia on average. 0:08:46.010,0:08:52.680 1.3 million times €100,[br]so we’re at €130 million 0:08:52.680,0:08:57.680 of economic welfare, of something that[br]people were willing to spend money for 0:08:57.680,0:09:01.350 that they’re not spending money for anymore[br]because of Wikipedia, because now that 0:09:01.350,0:09:05.930 we have Wikipedia most of the encyclopedias[br]aren’t actually useful for us anymore 0:09:05.930,0:09:10.100 because the knowledge that we have,[br]the knowledge that they would have 0:09:10.100,0:09:18.760 would be outdated very, very soon and[br]Wikipedia tends to be more up to date. 0:09:18.760,0:09:23.550 Well, that was from the consumer’s side.[br]But what about the business side? 0:09:23.550,0:09:29.319 There’s a lot of research on whether the[br]internet actually increases productivity 0:09:29.319,0:09:33.500 for businesses or not. Well, I don’t really[br]want to go into that debate because 0:09:33.500,0:09:38.209 it’s a really long tedious debate that is[br]kind of focused on “Well, you did this 0:09:38.209,0:09:42.110 method wrong”, or “You did this wrong”,[br]and “Well, I don’t think your argument 0:09:42.110,0:09:47.410 makes sense”. So it’s very… I don’t like[br]this kind of debate. I really like to go 0:09:47.410,0:09:51.720 deeper in things. But one of the things[br]that I found was that a lot of businesses 0:09:51.720,0:09:59.219 do rely on the internet by now. Now[br]we can see on this graph that most firms, 0:09:59.219,0:10:06.199 overall about 70% of firms actually[br]use the email to communicate. 0:10:06.199,0:10:09.550 Now email obviously only works[br]if you have internet, so they need 0:10:09.550,0:10:16.450 some sort of access to internet in order[br]for their current business model to work. 0:10:16.450,0:10:22.110 Now this was just some short ideas on[br]sort of what can the internet mean for 0:10:22.110,0:10:26.149 the economy. And now I want to talk about[br]Internet censorship, just a little bit. 0:10:26.149,0:10:33.620 Now, I’m not a censorship expert. I’m just[br]someone who read a lot of papers about it, 0:10:33.620,0:10:37.660 and who was very interested in what kind[br]of effects this has beyond sort of 0:10:37.660,0:10:43.890 the obvious “people don’t have access[br]to political information”. 0:10:43.890,0:10:47.709 So first a definition. ‘Internet censorship’[br]is the controller suppression 0:10:47.709,0:10:50.889 of what can be accessed, published[br]or viewed on the Internet 0:10:50.889,0:10:55.269 enacted by regulators or on their own[br]initiative. Now, in trying to conceptualize 0:10:55.269,0:10:59.269 internet censorship, for me, personally,[br]there’s two dimensions that are 0:10:59.269,0:11:03.840 very important. One is how targeted[br]is this internet censorship? 0:11:03.840,0:11:11.509 Now, you could, in theory, basically[br]have internet censorship 0:11:11.509,0:11:15.339 that is very, very targeted,[br]which you see in some cases. 0:11:15.339,0:11:18.729 Or you can have censorship[br]that isn’t targeted at all, like in Egypt. 0:11:18.729,0:11:23.560 They just decided to close the internet[br]down, basically, for a day. 0:11:23.560,0:11:28.249 That isn’t very targeted censorship,[br]obviously. The other thing to look at 0:11:28.249,0:11:32.980 is how widespread is it? So if you are[br]a business or if you’re a normal consumer 0:11:32.980,0:11:38.939 how probable is it that you would come (?)[br]something that’s censored? 0:11:38.939,0:11:43.180 Now, obviously, if you’re in China it’s[br]a lot more probable that you would 0:11:43.180,0:11:47.159 try to access something that’s censored[br]than if you’re in Germany. Even though 0:11:47.159,0:11:52.720 Germany also does some censorship.[br]And the way I like to conceptualize it is 0:11:52.720,0:11:58.019 to be kind of on a continuum. So I don’t[br]look… I don’t say “Well, either 0:11:58.019,0:12:01.649 there’s censorship or there isn’t[br]censorship”. What I’m trying to say is 0:12:01.649,0:12:06.930 “Censorship has a big spectrum[br]of things that can happen”. 0:12:06.930,0:12:12.809 These are some types of Internet censorship[br]that have different sort of implications. 0:12:12.809,0:12:16.309 I don’t want to go through them in detail[br]because I think we’ve heard some really 0:12:16.309,0:12:21.540 interesting talks on Internet censorship[br]already. But this is kind of 0:12:21.540,0:12:26.540 interesting or important for the model[br]that I’m trying to build. 0:12:26.540,0:12:30.370 But before trying to build my model,[br]first some more motivation. 0:12:30.370,0:12:33.980 I was trying to look at “is there any[br]evidence that it would have 0:12:33.980,0:12:39.819 an economic impact?”. And there actually[br]is a study that’s conducted by sort of 0:12:39.819,0:12:45.819 lobbying organizations, so obviously[br]should be taken with a grain of salt. 0:12:45.819,0:12:50.310 But it is quite interesting, and it shows[br]that there seems to be a correlation 0:12:50.310,0:12:59.499 between freedom and how good[br]the economic impact of internet is. 0:12:59.499,0:13:03.769 This is just a simple correlation. You can[br]see that there’s a really good line 0:13:03.769,0:13:09.920 going through it. They did do some[br]controlling for GDP per capita, so 0:13:09.920,0:13:16.580 for development level. But it still seems[br]quite rudimentary, to be honest. 0:13:16.580,0:13:23.979 The data that they use is quite bad[br]because it is very, very… 0:13:23.979,0:13:30.289 it’s just not finally granular enough, and[br]a lot of it is kind of… someone rating… 0:13:30.289,0:13:34.809 so “How do you think the economic…”,[br]“How do you think Internet 0:13:34.809,0:13:39.699 impacts the economy in this country?”[br]And then this is the data that they use, 0:13:39.699,0:13:48.069 to some degree. So it seemed very…[br]it didn’t really seem like a good, final answer. 0:13:48.069,0:13:53.529 So I’m trying to set up my own model.[br]And in my model I have a government 0:13:53.529,0:13:57.709 that chooses the type of censorship. And[br]for this type of censorship that it chooses 0:13:57.709,0:14:02.509 it pays a cost. Because we all know[br]censorship can be very expensive. 0:14:02.509,0:14:09.709 And in my model for now the only type of[br]expenses that I calculate are actual 0:14:09.709,0:14:16.989 manpower and technology expenses. I don’t[br]calculate reputation expenses at this point. 0:14:16.989,0:14:24.209 There is… there are firms in n industries.[br]Now this n is kind of not a fixed number 0:14:24.209,0:14:30.629 but instead is a number that can fluctuate[br]depending on the kind of country 0:14:30.629,0:14:37.879 I’m trying to model. And these industries[br]distinguish themselves by their 0:14:37.879,0:14:42.459 information intensity, or what I like[br]to call ‘information intensity’. Basically 0:14:42.459,0:14:47.540 I look at information as a commodity.[br]And what I’m trying to decide, or 0:14:47.540,0:14:51.910 the way I distinguish different kinds of[br]industry is how important is information 0:14:51.910,0:14:56.279 as a commodity, as opposed to other kinds[br]of commodities that are important 0:14:56.279,0:15:01.160 for this industry. So let’s look at[br]information intensity equals Zero. 0:15:01.160,0:15:05.259 Like if we don’t really… if information[br]as a commodity really isn’t important, 0:15:05.259,0:15:09.720 especially sort of conveyed information,[br]transmitted information. We can 0:15:09.720,0:15:14.309 think of traditional agriculture. Now[br]I know today’s agriculture tends to be 0:15:14.309,0:15:18.859 large-scale, and there’s a lot of[br]technology involved. But if you look at 0:15:18.859,0:15:24.170 very traditional agriculture that we[br]still might see happening in some parts 0:15:24.170,0:15:30.060 of Africa there usually is very, very[br]little information transmission involved. 0:15:30.060,0:15:34.069 And most of the information transmission[br]that is involved is actually mostly through 0:15:34.069,0:15:40.189 word of mouth. So that would be a case of[br]information intensity of very close to Zero. 0:15:40.189,0:15:43.790 And then if we look at information intensity[br]of 1 where basically the internet is 0:15:43.790,0:15:48.759 the most… or information is the most[br]important commodity. Internet businesses 0:15:48.759,0:15:54.839 themselves would… obviously qualify here,[br]– sorry – like, let’s look at Facebook 0:15:54.839,0:15:59.899 and other kinds of businesses like this.[br]And in between we have sort of industrial 0:15:59.899,0:16:03.339 companies in the modern world.[br]Now if we’re closer to the Zero end 0:16:03.339,0:16:07.639 of the spectrum we might be[br]at 0.2 .. 0.3, something like this, 0:16:07.639,0:16:15.449 we might be in traditional garment[br]factories. They do have information needs, 0:16:15.449,0:16:20.720 they get their cuts and stuff from the[br]Internet by now, or by email. 0:16:20.720,0:16:25.129 But once they have them they basically stay[br]the same for a couple of weeks or months. 0:16:25.129,0:16:30.409 So there’s a very low information[br]requirement. On the other side, 0:16:30.409,0:16:35.999 closer to 0.8 or something[br]like that we have high-tech, 0:16:35.999,0:16:41.220 especially software manufacturing,[br]so to speak. Information and being able 0:16:41.220,0:16:44.930 to transmit this information is very[br]important. Now, in between we might look 0:16:44.930,0:16:51.259 at traditional industrial companies[br]like automobile manufacturing 0:16:51.259,0:16:56.000 that might be somewhere in between.[br]And before the game, or before… 0:16:56.000,0:17:00.160 or at the first run of the model[br]‘service level’ and ‘globalization level’ 0:17:00.160,0:17:05.599 are randomly distributed. The information[br]intensity of industries is also kind of 0:17:05.599,0:17:11.799 randomly distributed, but not in a true[br]random fashion. Because when looking 0:17:11.799,0:17:15.500 in the wild, sort of what kind of[br]economies exist, most of them… 0:17:15.500,0:17:19.199 the information intensity of one[br]industry is kind of correlated with 0:17:19.199,0:17:23.449 information intensities of other industries[br]in this country. Like in Germany 0:17:23.449,0:17:29.269 we’re very known for a certain type[br]of industry that we have quite a lot of, 0:17:29.269,0:17:35.440 which is manufacturing, very high-technology[br]manufacturing. So we have more industries 0:17:35.440,0:17:40.450 in this area but we have less traditional[br]agriculture, for example. 0:17:40.450,0:17:44.669 So having a true random distribution[br]wouldn’t work. In addition the service level 0:17:44.669,0:17:49.919 and the globalization level are randomly[br]distributed as kind of external variables. 0:17:49.919,0:17:55.090 Obviously, this is a simplification because[br]I can’t really start at the beginning like 0:17:55.090,0:17:58.870 I can’t say: “Oh well, I’ll start,[br]I don’t know, 2000 BC 0:17:58.870,0:18:04.190 with a very blank economy, and then[br]something happens and something happens 0:18:04.190,0:18:08.320 and something happens”. That’s just not[br]realistic. So in order to get a better idea 0:18:08.320,0:18:12.830 of what happens with different types of[br]economies, what I’m doing is I’m running 0:18:12.830,0:18:18.899 this game or this model again and again.[br]And having these random parameters 0:18:18.899,0:18:24.539 basically changed everytime.[br]So on average there should be… 0:18:24.539,0:18:29.289 there should be usable results. 0:18:29.289,0:18:35.230 Now what this is actually missing[br]is the consumer as a labourer. 0:18:35.230,0:18:40.090 So I don’t really have ‘labour’ reflected[br]in here. A more complete model would have 0:18:40.090,0:18:44.080 that reflected. But it’s not the most[br]interesting aspect of my model, so 0:18:44.080,0:18:49.940 I’m not presenting this here, basically. 0:18:49.940,0:18:56.080 Now, let’s look at what this would[br]mean for firms. In my model 0:18:56.080,0:18:59.649 what kind of things would I expect[br]thinking through it logically which is 0:18:59.649,0:19:04.820 always the first step when trying to model[br]something. First of all if we have 0:19:04.820,0:19:10.130 an information intensity of something[br]greater than Zero but smaller than One. 0:19:10.130,0:19:14.410 Because the information intensity being[br]close to One is kind of a special case 0:19:14.410,0:19:18.520 that I’ll be talking about later on.[br]Internet censorship increases the cost 0:19:18.520,0:19:22.360 and uncertainty of information.[br]And of course that is more important 0:19:22.360,0:19:27.850 the more important information is[br]for this certain industry. 0:19:27.850,0:19:33.850 So for a traditional garment factory[br]internet censorship might be a lot 0:19:33.850,0:19:41.000 less important than for a semiconductor[br]factory that has to receive 0:19:41.000,0:19:47.090 new blueprints every day or every month[br]or something. The second thing is 0:19:47.090,0:19:51.559 the more globalized the economy as a whole[br]is the more costly internet censorship 0:19:51.559,0:19:58.490 will be. Similar reasoning. 0:19:58.490,0:20:02.990 And another thing for firms is the[br]less focused the censorship 0:20:02.990,0:20:07.640 the higher the cost. Now this assumes that[br]the censorship or the goal of censorship 0:20:07.640,0:20:14.370 usually isn’t to turn down firms or to[br]make sure that firms don’t succeed. 0:20:14.370,0:20:19.820 So if censorship is very focused[br]firms tend to be affected less 0:20:19.820,0:20:25.149 which makes their associated cost less.[br]Now of course we can argue, well, 0:20:25.149,0:20:29.399 firms can circumvent censorship, and they[br]can do that for sure. But it is expensive 0:20:29.399,0:20:35.299 to do that. If you’ve ever tried a VPN[br]in China e.g., first, buying the VPN 0:20:35.299,0:20:40.919 is expensive. Then, having someone sort of[br]make sure that the VPN works is expensive, 0:20:40.919,0:20:44.009 every couple of months you need to change[br]it because the Chinese Government decides, 0:20:44.009,0:20:52.549 well, this VPN shouldn’t work anymore. So[br]it’s a very expensive and uncertain thing, 0:20:52.549,0:20:57.909 really. For firms in[br]‘information intensity = 1’ 0:20:57.909,0:21:02.940 it obviously also increases the cost[br]of operating. Some of these firms actually 0:21:02.940,0:21:07.970 carry out some censorship for governments.[br]We have seen that happening more recently. 0:21:07.970,0:21:12.570 But there might actually be some firms[br]that have a relative advantage, especially 0:21:12.570,0:21:16.820 domestic firms often have a relative[br]advantage due to the censorship because 0:21:16.820,0:21:20.950 they know the regulators better, they know[br]how to deal with it, they might have 0:21:20.950,0:21:25.039 less need to circumvent, actually.[br]And even if they do need to circumvent 0:21:25.039,0:21:29.539 it’s easier for them because[br]they speak the language etc. 0:21:29.539,0:21:34.090 This is actually a special case that I’ll[br]be talking about a little bit later as well. 0:21:34.090,0:21:38.460 For the government – I’ve said[br]that censorship is costly. But moreover, 0:21:38.460,0:21:43.100 the more targeted and accurate censorship[br]is the more manpower and technology intensive 0:21:43.100,0:21:50.389 it actually is. This is a finding by[br]Leberknight et al. in a research paper. 0:21:50.389,0:21:54.480 I think they’re electrical engineers, and[br]they calculated through different types 0:21:54.480,0:22:00.350 of censorships and how expensive it would[br]be to scale them up. So that is actually 0:22:00.350,0:22:03.479 a really interesting finding because[br]it shows that for governments 0:22:03.479,0:22:10.460 having sort of less targeted censorship[br]is less costly. But this is the kind of 0:22:10.460,0:22:17.039 censorship that is actually most affecting[br]in a negative way to firms, 0:22:17.039,0:22:20.990 in an economy. So that’s kind of not[br]a result that we would really want 0:22:20.990,0:22:24.919 because the incentives don’t line up in[br]that way. And economists love to talk 0:22:24.919,0:22:29.169 about incentives, obviously. Now for[br]consumers, they would obviously get 0:22:29.169,0:22:33.090 less benefits through the internet, the[br]benefits that I’ve talked about before. 0:22:33.090,0:22:38.430 And also businesses often pass on the cost[br]to consumers. 0:22:38.430,0:22:43.350 Now however, some countries[br]still benefit from internet censorship. 0:22:43.350,0:22:45.970 I’ve talked mostly [br]about why it’s costly to do it, 0:22:45.970,0:22:48.700 and I think it is costly in most cases. 0:22:48.700,0:22:53.210 But developing countries that start out at[br]low service and low globalization levels 0:22:53.210,0:22:58.950 usually have… in these kind of situations[br]internet censorship has less of an impact, 0:22:58.950,0:23:04.370 less of a negative impact.[br]And censorship can actually act 0:23:04.370,0:23:08.880 as protectionism. In information intensive[br]industries governments can use this kind 0:23:08.880,0:23:13.650 of censorship to push domestic industries[br]and enable catch-up growth. Now there 0:23:13.650,0:23:16.820 are a couple of further prerequisites.[br]First of all, the country needs to be 0:23:16.820,0:23:20.640 large enough so that these [br]information intensive industries 0:23:20.640,0:23:23.640 have a domestic market as well. 0:23:23.640,0:23:27.379 Obviously. And then also only[br]targeted censorship can serve as 0:23:27.379,0:23:32.159 protectionism. The only other way would be[br]if you decided on a domestic intranet and 0:23:32.159,0:23:38.059 basically closed your entire intranet off[br]to the world. Which is kind of difficult. 0:23:38.059,0:23:41.850 But what about the long-term effects[br]of that? Would they still be positive 0:23:41.850,0:23:47.669 for the government? Now, I’m using[br]‘positive’ in a very… sort of something 0:23:47.669,0:23:51.820 that should be taken with a grain of salt,[br]obviously. And what I did is I looked 0:23:51.820,0:23:57.330 at China. Obviously, I’m a China watcher.[br]So I’m really interested in China. And 0:23:57.330,0:24:02.190 this is kind of where my interest started.[br]And I’m really trying to find a framework 0:24:02.190,0:24:07.219 where China isn’t the exception but[br]instead China kind of fits into the model. 0:24:07.219,0:24:13.129 What we see is the Chinese government has[br]outsourced much if its censorship to these 0:24:13.129,0:24:19.000 internet companies. Baidu, Sina weibo,[br]Tencent probably would not exist by now, 0:24:19.000,0:24:24.820 actually, if the censorship didn’t exist.[br]And what we actually see now is that 0:24:24.820,0:24:29.750 WeChat e.g. is going global. It has[br]more functionality than Whatsapp 0:24:29.750,0:24:35.799 and they’re trying to get out. But as I’ll[br]be talking about later on a little bit 0:24:35.799,0:24:41.810 the censorship is starting to be a problem[br]for these companies that used to benefit. 0:24:41.810,0:24:46.840 There’s some things about Chinese… about[br]the character of Chinese Internet censorship 0:24:46.840,0:24:54.409 that is relevant here. But what about[br]the future? Now first it’s difficult to 0:24:54.409,0:24:58.660 innovate with this kind of censorship. And[br]this kind of insular education that we see 0:24:58.660,0:25:03.450 also makes innovation, real innovation,[br]very difficult. In China e.g. Github 0:25:03.450,0:25:07.631 is blocked most of the time. That makes[br]kind of collaborating, especially in 0:25:07.631,0:25:11.730 coding environments, very, very hard.[br] 0:25:11.730,0:25:14.490 Second, we see more global internet enabled 0:25:14.490,0:25:20.059 supply chains in the world. So if we have[br]these global Internet-enabled supply chains 0:25:20.059,0:25:25.669 having internet censorship turns out to be[br]more of a disadvantage the more globalized 0:25:25.669,0:25:31.879 these supply chains actually become. And[br]information becomes the most important 0:25:31.879,0:25:36.230 commodity all throughout China. Now this[br]of course also makes Internet censorship 0:25:36.230,0:25:41.000 more costly for the economy. What about[br]possible positives? So what could work 0:25:41.000,0:25:45.500 in the Chinese government’s favour? First,[br]the Chinese intranet is actually pretty 0:25:45.500,0:25:50.429 attractive to most people. Most people[br]don’t try to go outside, even like 0:25:50.429,0:25:55.269 they don’t even know that they can’t. They[br]just don’t want to do it. Second, the IoT, 0:25:55.269,0:25:59.429 where machines communicate with each other[br]doesn’t need to be affected because 0:25:59.429,0:26:04.820 most of the censorship that we see[br]happening could be reworked in a way 0:26:04.820,0:26:08.599 that doesn’t affect machine-to-machine[br]communication. And that wouldn’t be 0:26:08.599,0:26:14.039 a problem for what the censorship intends[br]to do which is sort of suppress political 0:26:14.039,0:26:20.669 opposition. And a third, the government[br]wants an economy more focused on domestic 0:26:20.669,0:26:24.230 consumption. So if they want to do this[br]then censorship might actually be good 0:26:24.230,0:26:30.669 for that. Now, for me, what I found out[br]when doing this research is first, 0:26:30.669,0:26:34.709 standard economic models really aren’t[br]suited for this kind of question. Because 0:26:34.709,0:26:38.370 they tend to use GDP, and I’ve told you[br]why GDP really is not a good measure 0:26:38.370,0:26:43.419 for that. Second, the next step that[br]I’ll be doing is agent-based modeling. 0:26:43.419,0:26:48.910 But I would really like to feed my models[br]with some reliable data. And I can’t 0:26:48.910,0:26:53.400 really find any of that. I can find some[br]data going back a couple of years 0:26:53.400,0:26:57.779 on, like, is there censorship, is there[br]no censorship. But I can’t really find any 0:26:57.779,0:27:02.150 good data that distinguishes between[br]different types of censorship, which would 0:27:02.150,0:27:06.440 be really important for the kind of[br]research that I really want to carry out 0:27:06.440,0:27:11.610 in the future. Thank you, guys. If you[br]have questions you can ask now or 0:27:11.610,0:27:15.129 you can come to me later, you can[br]of course also send me an e-mail. 0:27:15.129,0:27:18.719 I’m always happy to talk about this topic. 0:27:18.719,0:27:27.529 applause 0:27:27.529,0:27:32.000 Herald: Thank you very much for this talk.[br]We have six microphones at the floor level 0:27:32.000,0:27:35.660 here, so if you have questions we have[br]a very brief amount of time. 0:27:35.660,0:27:40.430 Please line up at the microphones.[br]We have microphone no. 2 over here. 0:27:40.430,0:27:46.480 Question: I want to mention one thing.[br]Always when talking about China censorship 0:27:46.480,0:27:51.299 this censorship applies to China main[br]land. So it’s not Hong Kong and not Taiwan. 0:27:51.299,0:27:51.959 Toni: Yes. 0:27:51.959,0:27:55.769 Question: And my question I want [br]to ask is: 0:27:55.769,0:27:59.219 What do you think about productivity [br]of work? 0:27:59.219,0:28:05.200 So e.g. if you shut down Facebook do you[br]think this would increase working 0:28:05.200,0:28:08.059 productivity?[br]Toni laughs 0:28:08.059,0:28:13.010 applause[br]Toni: That’s a really interesting question, 0:28:13.010,0:28:16.470 and something that I haven’t seen anywhere[br]in literature. There is a big literature 0:28:16.470,0:28:21.970 discussion about what the internet as such[br]means for productivity, and that’s 0:28:21.970,0:28:26.820 kind of both ways. Now, one of the things[br]to look at is that just because you 0:28:26.820,0:28:31.200 shut down Facebook doesn’t mean you[br]shut down any sort of social network. 0:28:31.200,0:28:36.389 And I do think that if people use Facebook[br]and suddenly aren’t able to use it anymore 0:28:36.389,0:28:40.769 they would probably spend their resources[br]trying to find new ways to access Facebook 0:28:40.769,0:28:48.790 which would probably not exactly[br]improve their productivity. 0:28:48.790,0:28:52.299 Herald: Next question[br]from microphone no. 2. 0:28:52.299,0:28:57.909 Question: Would it make sense to have[br]a model where firms use information 0:28:57.909,0:29:02.480 as an input to a production function and[br]then model censorship as a kind of tax 0:29:02.480,0:29:08.109 on that. That will seem like standard new[br]classical micro-econ one-on-one stuff? 0:29:08.109,0:29:12.390 Toni: That would make sense. I’ve actually[br]looked at this. One of the problems with 0:29:12.390,0:29:17.730 doing that is that information [br]as a commodity 0:29:17.730,0:29:23.350 is very difficult to be used in this new[br]classical way because you usually assume 0:29:23.350,0:29:28.020 that everything is kind of friction-less.[br]And if things are friction-less then 0:29:28.020,0:29:31.619 information can’t really be a commodity[br]because you assume that information 0:29:31.619,0:29:36.500 basically gets transferred immediately,[br]and without any sort of censorship. So 0:29:36.500,0:29:39.590 we can talk about this a little bit later.[br]Maybe you have some ideas that 0:29:39.590,0:29:43.740 I haven’t found yet.[br]It would be interesting. 0:29:43.740,0:29:47.539 Herald: And the next question,[br]as well, from microphone no. 2. 0:29:47.539,0:29:53.629 Question: So, going the same direction:[br]for GDP is rather defined what is 0:29:53.629,0:29:59.429 the optimization problem for a government.[br]For your further approaches what would be 0:29:59.429,0:30:05.279 the optimization that a government like[br]China does then. If you say e.g. Wikipedia 0:30:05.279,0:30:08.950 which leaks out to all over the world but[br]what is the government optimizing then? 0:30:08.950,0:30:15.049 Toni: What I’m looking at is economic welfare[br]as defined as producer and consumer surplus. 0:30:15.049,0:30:22.539 And I assume that the government’s goal[br]is to optimize economic welfare for both 0:30:22.539,0:30:27.519 producers, consumers and also for itself[br]as a producer and as a consumer. 0:30:27.519,0:30:32.240 Question: So your criticism is more like[br]you don’t have a good proxy, 0:30:32.240,0:30:33.870 using GDP for economic welfare? 0:30:33.870,0:30:36.870 Toni: Yes, yes.[br]Okay. Thank you. 0:30:36.870,0:30:38.370 Herald: I’m afraid we’re all out of time. 0:30:38.370,0:30:40.350 Please give a warm round[br]of applause to Toni! 0:30:40.350,0:30:43.690 applause 0:30:43.690,0:30:46.260 post-roll music 0:30:46.260,0:30:50.540 Subtitles created by c3subtitles.de[br]in the year 2017. Join, and help us!